A1529
Title: Robust synthetic control method for data with outliers
Authors: Riku Yamashita - Doshisha University (Japan) [presenting]
Kei Tsubotani - Graduate School of Doshisha University (Japan)
Kensuke Tanioka - Doshisha University (Japan)
Hiroshi Yadohisa - Doshisha University (Japan)
Abstract: In practical applications, accurately estimating the effects of interventions on outcomes is important. The synthetic control method is widely applied in various fields, such as economics and political science, particularly when randomized experiments are not feasible. This method produces a potential outcome that would have been observed for the treated unit in the absence of treatment by creating a weighted combination of control units that closely matches the characteristics of the treated unit before the intervention. However, the standard synthetic control approach is sensitive to outliers in the treated unit, leading to biased estimates. To address this problem, an outlier-robust synthetic control method that replaces the L2 norm is proposed in the objective function with the L1 norm. This approach can reduce the influence of outliers and improve the robustness of the proposed method. This modification enhances the accuracy of the treatment effect, and the applicability of the synthetic control method may extend to a wider range of real-world scenarios. The effectiveness of the proposed method is demonstrated through simulations and empirical examples, highlighting its practical utility in contexts where data anomalies or outliers are present.
